Software Engineer - Traffic Capacity & Performance

Meta's edge network carries traffic in two directions: content out to billions of people from edge sites and caches inside internet service providers, and Meta's own services and AI agents reaching sites across the internet. The Traffic Capacity & Performance team decides how much capacity Meta builds for both, and where. We size it to keep serving through failures and demand spikes, drain sites at peak on purpose to find weak spots before a real outage does, and have apps and services ask for less when capacity runs low.

Responsibilities

Build and extend large-scale simulation of global traffic placement: how traffic shifts across edge sites as demand grows and as sites, peering links and backbone paths fail. Run thousands of scenarios fast enough to plan with
Work with the load-balancing team on simulation fidelity and performance, and validate results against production behavior and live failure tests
Formulate and solve capacity planning as constrained optimization: where to add compute, peering and caches, given space, power, supply, latency and failure constraints. Build and tune the solvers
Plan capacity for the browsers, proxies and caches Meta's services and AI agents use to reach the internet, and find which one becomes the bottleneck as Meta's internet access needs grow
Build production forecasting of peak demand, by region and by product, driven by product changes and events rather than trend alone. Quantify uncertainty and track accuracy against actuals
Model failures and their impact, including correlated and cascading failures. Size the buffers we hold against failures and demand surges, and express plans as risk and cost trade-offs leadership can choose between
Build agentic planning systems that rerun forecasts and plans continuously, detect drift, explain what changed and propose actions, with evaluations and guardrails
Lead technical direction across teams, mentor engineers, and represent capacity engineering in cross-team design reviews
Run what you build in production, including oncall

Qualifications

Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
7+ years of software development experience in industry, or a PhD with 4+ years of experience
Proficiency in Python and at least one of C++, Java, Go or Rust
Experience building and operating production systems or large-scale data pipelines
Experience in at least one of: large-scale simulation, mathematical optimization, or forecasting systems
Experience leading technical work that spans multiple teams Experience running large scenario sweeps on distributed compute (for example Spark or Presto)
Experience building discrete-event, flow-level or Monte Carlo simulations, and validating them against production
Experience applying LLM-based agents to production workflows, including evaluations
Experience with mixed-integer or linear programming and solvers (for example Gurobi, CPLEX, OR-Tools or HiGHS), or large-scale heuristic search
Experience with network capacity planning, traffic engineering, content delivery networks, load balancing or forward proxies at scale
Experience with reliability engineering: failure modeling, failure-domain analysis, disaster-recovery or chaos testing
Experience with probabilistic, driver-based forecasting, including launches and events
Advanced degree in operations research, computer science, electrical engineering, applied mathematics or a related field